I am thoroughly confused again. I am at the point of relegating electricity as a black box, a mysterious phenomena I'll never comprehend. I just saw a YouTuber build an elaborate water mill for generating electricity. The water wheel spins a standard 3-phase squirrel cage motor wired to 3 capacitors (Linked a few seconds of another video strictly for reference). I didn't see a single comment questioning his generator, leading me to question my own understanding of induction motors and electricity in general.

From my simple understanding, a squirrel cage induction motor doesn't use permanent magnets either on the rotor or the stator for inducing current on the windings. Instead, alternating current is applied to the windings to produce a rotating magnetic field which mechanically spins the rotor. In principle, it converts electric energy into mechanical energy, not the other way around.

If my understanding is correct, then a squirrel cage induction motor cannot convert mechanical energy into electrical energy. Unlike a DC motor, it has no permanent magnets and relies on purely on electromagnetism. Spinning the rotor won't induce current in the stator's winding since the rotor doesn't have a magnetic field.

It could be argued the stator windings could be energized to induce a magnetic field to induce current on the rotor. However, the rotor's bars are shorted, and electrically isolated unlike a DC motor uses brushes with no output path.

There's also possibility that the term induction is applied liberally to motors and generators that are not squirrel cage motors.

Is my understanding incorrect?

Video demonstration of an induction motor running as a generator


  • \$\begingroup\$ Does This is a question about squirrel-cage induction machine theory help? \$\endgroup\$ Commented Dec 16, 2023 at 20:11
  • \$\begingroup\$ @ChesterGillon Unfortunately, it's not terribly helpful. I will add further clarification to my question. Specifically about the rotor. \$\endgroup\$
    – user148298
    Commented Dec 16, 2023 at 20:17
  • \$\begingroup\$ How about en.wikipedia on induction generators? \$\endgroup\$
    – greybeard
    Commented Dec 16, 2023 at 20:55
  • \$\begingroup\$ Great link greybeard! Even has a worked example for the capacitors. How did we ever do anything before the interwebs!! \$\endgroup\$
    – Kartman
    Commented Dec 17, 2023 at 0:58
  • \$\begingroup\$ @greybeard I had already read through it and it didn't address my question for the same reason I already mentioned. \$\endgroup\$
    – user148298
    Commented Dec 19, 2023 at 1:58

2 Answers 2


The Wikipedia article, at time of writing, seems to contain an error:

The generating mode for induction motors is complicated by the need to excite the rotor, which being induced by an alternating current is demagnetized at shutdown with no residual magnetization to bootstrap a cold start. It is necessary to connect an external source of magnetizing current to initialize production. The power frequency and voltage are not self regulating.

[my emphasis]. Arguably this construction is of the type "here are assumptions X, from which we conclude Y", and the error is merely in not making the assumptions contingent ("if assumptions X, then we conclude Y").

This arguably might even be more of a style choice than an outright miscommunication, as listing ones' assumptions indicates to the reader what conditions can contradict the conclusion. But it's not a generous or clear style, as novice readers, who aren't yet comfortable enough with a given subject, won't know when something is a postulate or an assumption, or when to question then (and which ties in with the general subject of critical thinking).

Anyway, my understanding of it is, either due to residual magnetization in the rotor, random electrical noise, or the presence of a starting circuit, and in combination with capacitors to cancel out the stator's inductance, both startup and continuous operating requirements are met, and generation is had.

Mind, I haven't built or installed any such generators myself, so this is hearsay (notice my lack of citations here). From my knowledge of magnetic materials (silicon steels have pretty generous remanence, though that will come down a fair bit when including the effect of air gap, coercivity, and ringdown, but nonzero still does not equal "zero"), it sounds highly plausible.

It probably also matters if the load is small during startup, so it doesn't need to regenerate field into an electrical load at the same time.

As for a mechanical explanation, note that the torque curve is symmetrical around zero slip, so when generating at a given torque, the output frequency slips higher than the rotational frequency, by the same amount it would drift below due to an equal loading torque instead.

It is also important to note that the reactive current of an induction motor is roughly constant (independent of torque, plus or minus), and mainly the active power (real current) varies with load, obviously being proportional around zero (load current for loading torque, generated current for generating torque).

The reactive current is a consequence of stator inductance (essentially the inductance due to the air gap, assuming the rotor is a solid hunk of electrical steel with no squirrel cage; which is what it looks like when it is spinning synchronous to the stator field i.e. at zero slip), plus the in-phase component of the rotor's field with respect to the stator (in the rotating reference frame, which is normally equal by the same reason).

When excitation is provided, the consequence either way is simply an inductive current draw, from the constant-voltage, fixed-frequency source; when self-generating however, the reactive current impresses a phase and thus angle shift on the induced rotor current, damping it out over time. If enough capacitance is connected, to at least cancel out the inductive current, then the loading phase angle acts to advance the rotor's induced current, building it up over time instead.

Thus, the load must remain capacitive as seen by the generator, otherwise a neutral or lagging current phase causes the rotor current to decay over time. A capacitive load gives rising rotor current, up to the limit determined by magnetic saturation of the rotor and/or stator.

Since (stand-alone) output is defined by slip and magnetic saturation, not only is the frequency fairly poorly defined (load dependent), but voltage is frequency (and thus load) dependent as well. Normally, motors are designed quite close to saturation, so the generator-mode output of a common induction motor isn't much more than the nameplate rating, and they are usable this way, if a bit crude.


a squirrel cage induction motor doesn't use permanent magnets either on the rotor or the stator for inducing current on the windings Correct.
Same as big alternators or ("true") dynamos.

it converts electric energy into mechanical energy yes…
not the other way around now that's just to do with the designation of an induction machine as a motor. You can name one and the same induction machine an induction generator when driving it with mechanical power to "negative" slip, oversynchronous speed, when grid operated.

(I find grid operation more intuitive than stand-alone, possible (only) with capacitors providing for reactive power.)


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.